Abstract:Humans can solve image classification tasks by learning from a few images and reusing prior-knowledge. In Artificial Intelligence, deep-learning models have been implemented to simulate human learning and face problems with little data available, few-shot learning. Nevertheless, one crucial problem of deep-learning is the selection of architectures and initial parameters that accomplish the requirements for a specific task. Therefore, we propose a novel methodology based on the meta-learning paradigm, which reuses prior- and meta-knowledge to select the best architecture and its initial parameters to solve different few-shot image classification tasks, as humans do. Our methodology was designed to understand the knowledge flow of the meta-learning paradigm by dividing the learning into prior-models, meta-models, meta-unsupervised algorithm, and base-models. We considered 9 architectures of deep convolutional neural networks as prior-models. Also, we propose a meta-unsupervised algorithm inspired by the human-cognitive problem-solving process, which acquires knowledge by solving tasks to recommend initial parameters to solve other new tasks. Furthermore, we propose a New Task Distribution Scheme to better evaluate few-shot learning models and analyze the difficulty of the new tasks. We evaluated our meta-learning methodology by applying the NTDS to the Mini-ImageNet and Caltech-UCSD-Birds-200-2011 (CUB) datasets. The achieved Adjusted Mutual Information average scores are 0.955 (zero-shot) and 0.957 (five-shot) for Mini-ImageNet, and 0.822 (zero-shot) and 0.848 (five-shot) for CUB. Moreover, we demonstrated the applicability of our meta-learning methodology in a real-world scenario using a new copro-parasite dataset. The performances of our meta-learning methodology are competitive regarding the state-of-the-art unsupervised few-shot learning models.

The meta-learning method for the ensemble model based on situational meta-task

Meta-Learning across Meta-Tasks for Few-Shot Learning

Revisiting Unsupervised Meta-Learning via the Characteristics of Few-Shot Tasks

Prior-Knowledge and Attention-based Meta-Learning for Few-Shot Learning

Hierarchical Meta-Learning with Hyper-Tasks for Few-Shot Learning

A metric-based meta-learning approach combined attention mechanism and ensemble learning for few-shot learning

Task Attended Meta-Learning for Few-Shot Learning

Rethinking Meta-Learning from a Learning Lens

Meta-learning methodology based on meta-unsupervised algorithm for meta-model selection to solve few-shot base-tasks

Set-based Meta-Interpolation for Few-Task Meta-Learning

Towards Well-Generalizing Meta-Learning Via Adversarial Task Augmentation

A Joint Optimization Framework of the Embedding Model and Classifier for Meta-Learning

A Novel Hierarchical Adaptive Feature Fusion Method for Meta-Learning

Meta Omnium: A Benchmark for General-Purpose Learning-To-Learn

Rethink and Redesign Meta Learning

EEML: Ensemble Embedded Meta-Learning

Revisiting Meta-Learning As Supervised Learning

Representation Based and Attention Augmented Meta Learning

Meta Learning to Bridge Vision and Language Models for Multimodal Few-Shot Learning

Few-Shot Learning of Compact Models via Task-Specific Meta Distillation